基于粒子群优化SVM的苹果图像分割

Application of SVM Algorithm for Particle Swarm Optimization in Apple Image Segmentation

苹果图像分割是苹果采摘机器人视觉系统中识别和定位的关键技术.针对目前苹果采摘机器人对果实识别误差大、处理时间长等问题,结合粒子群算法在求解组合优化问题时具有的全局搜索特性,提出了一种基于粒子群参数优化的SVM分割算法.试验结果表明:该算法能很好地实现苹果果实与图像背景的分离,后续利用数学形态学中的闭运算对分割后的图像进行处理,能够较好地保存苹果轮廓信息、消除孔洞现象,为完善苹果采摘机器视觉系统的识别和定位提供技术支持.

The apple image segmentation is the key technology of identification and location in the apple-picking machine vision system. On account of huge errors in the process of discriminating fruits by apple-picking robots at present and the long-time processing, the SVM theory in fingerprint image segmentation method is conducted. Combined with the global search ability of particle swarm optimization in solving combinational optimization problems, the SVM partitioning algorithm, which is based on the parameter optimization of particle swarm, is put forward. The results show that this algorithm makes the separation of apple fruits and the image background come true. It also preserves the outline of apples, then polishes the image after segmentation by the close operation in mathematical morphology, which eliminates the pore phenomenon effectively and provides convenience for the further apple-picking and apple-discriminating.